Reserved Instance
Reserved instances discount cloud compute in exchange for a multi-year usage commitment, best suited to steady, predictable workloads.
A reserved instance (RI) is a billing commitment that lowers the cost of cloud compute in return for reserving capacity over a fixed term, typically one or three years. Rather than paying on-demand rates, the customer commits to a certain usage and receives a substantial discount. Reserved instances suit workloads with steady, predictable demand.
How It Works
The customer selects parameters such as instance family, region, and term, and optionally pays all, part, or none of the cost upfront. The deeper the commitment and the larger the upfront payment, the bigger the discount. Many providers have evolved this concept into more flexible savings plans that commit to a dollar-per-hour spend rather than a specific instance type, letting the discount apply across instance families and even services. Reservations apply automatically to matching usage; unused reservations are wasted spend, so accurate forecasting matters.
Why It Matters
Reserved instances and savings plans are central to controlling cloud costs for baseline capacity. A typical strategy covers predictable, always-on load with reservations or savings plans, handles variable load with on-demand capacity, and runs interruptible work on cheaper spot instances. The risk is over-committing to capacity that demand no longer needs, which is why teams analyze utilization before purchasing and revisit commitments as workloads change.
Related Terms
Reserved instances complement spot instances and on-demand capacity and are a foundational FinOps practice for cloud infrastructure.